Did the Patriot missile really intercept 41 of 42 Scuds during the Gulf War?

That was President George H.W. Bush's public claim on February 15, 1991, while the war was still underway, according to the official transcript in the Public Papers of the Presidents. It did not hold up under later review: the Army's own postwar estimate came in lower, and a 1992 General Accounting Office investigation found only 9 percent of Patriot-Scud engagements supported by the strongest evidence that a warhead had actually been destroyed. (GovInfo, Public Papers of the Presidents; U.S. GAO, NSIAD-92-340)

What did the GAO conclude about Patriot's actual performance in the Gulf War?

In a September 1992 report, the General Accounting Office found that only 9 percent of Patriot-Scud engagements were backed by the strongest available evidence, physical debris or radar data confirming a warhead had actually been destroyed or disabled. For the remaining engagements, GAO concluded the Army could prove only that Patriots had come close to the Scuds, not that they had destroyed them. Raytheon publicly disputed the findings as based on incomplete data. (U.S. GAO, NSIAD-92-340; PBS Frontline, Raytheon's Response)

How many Americans died in the Dhahran barracks Scud attack, and why did Patriot fail to stop it?

Twenty-eight U.S. soldiers were killed on February 25, 1991, when a Scud struck an Army barracks in Dhahran, Saudi Arabia, the deadliest single attack on American forces of the war. The Patriot battery assigned to the site never tracked the incoming missile because of a software bug: after running continuously for more than 100 hours, an accumulating clock-timing error threw off its tracking by roughly a third of a second. A corrected software version existed but reached that battery on February 26, one day after the attack. (U.S. GAO, IMTEC-92-26)

How many Scud missiles did Iraq fire during the Gulf War?

Researchers have documented roughly 88 Scud launches between January 18 and February 25, 1991, with 46 aimed at Saudi Arabia and other coalition targets in the Kuwait theater and 42 at or near Israel. Iraq itself told United Nations inspectors after the war that it had launched 93 missiles, 50 against coalition targets and 43 against Israel, a discrepancy researchers attribute mainly to missiles that failed early in flight or landed inside Iraqi territory. (Wikipedia, Iraqi ballistic missile attacks on Saudi Arabia)

Gulf War Desert Storm: The 41-for-42 Claim and the 9 Percent the GAO Could Actually Confirm

President Bush declared the Patriot missile 41-for-42 against Iraqi Scuds three weeks into the war. Congress's own investigators could later verify a warhead kill in roughly one engagement out of eleven.

A Radar System Built for a Different War

The Patriot missile system entered U.S. Army service in 1984 as an anti-aircraft weapon, built to shoot down jets and cruise missiles, not the short-range ballistic missiles Iraq possessed. It traces back to 1965, when the Army began looking for a long-range replacement for its aging Nike Hercules and Hawk air-defense missiles; Raytheon won the development contract in 1967 for what was then called SAM-D, and the program was renamed Patriot in 1976 as it moved into full-scale development. A late-1980s software upgrade gave it a limited, largely untested capability against incoming ballistic missiles like the Scud, and that capability had never been used in actual combat before Iraq invaded Kuwait in August 1990. When Iraq began firing Scuds at Saudi Arabia and Israel on January 18, 1991, the day after the coalition's air campaign opened, Patriot's real debut arrived at the tail end of a decade in which whether any missile-defense system could reliably work at all had already been argued over, in a different Cold War context, as fiercely as the missiles themselves.

Israel had no Patriot batteries or trained crews of its own, and U.S. Army units deployed with their own batteries to defend Israeli cities within days of the first strikes, a response widely understood at the time as an effort to keep Israel from retaliating and fracturing the Arab states inside the U.S.-led coalition. The Army credited Patriot with its first combat intercept the opening night over Dhahran, Saudi Arabia, and claimed nine more Scuds shot down over Saudi cities and installations on the night of January 20. Those early, unverified claims, repeated on live television as Patriots streaked upward to meet incoming Scuds, set the frame for how the system's performance would be judged for the rest of the war.

Bush's 41-for-42 Claim at the Raytheon Plant

On February 15, 1991, with the war not yet at its halfway point and the ground offensive still nine days from launching, President Bush traveled to Raytheon's Patriot manufacturing plant in Andover, Massachusetts. According to the official transcript in the Public Papers of the Presidents, Bush told the assembled workers, "Patriot is 41 for 42: 42 Scuds engaged, 41 intercepted," putting the system's claimed success rate above 97 percent at a moment when the fighting, and the accounting of what Patriot had actually done, was far from finished.

That figure was never subjected to independent verification before Bush cited it, and it did not match what the Army itself would settle on once the war ended and analysts began reviewing radar tapes and recovered missile debris rather than relying on operators' real-time judgment calls made under fire.

The Army's Own Numbers Started Shrinking

In its postwar assessment, the Army settled on a success rate of 80 percent in Saudi Arabia and 50 percent in Israel, already well below the 97-percent figure Bush had cited at Raytheon. Those numbers did not hold either. Under pressure from congressional hearings scrutinizing the claims, the Army revised its own estimate downward again by April 1992, to 70 percent in Saudi Arabia and 40 percent in Israel.

A round, quotable number that circulates publicly without matching what the underlying paperwork actually supports is a pattern this war's own record has produced more than once, including in a Medal of Honor citation now popularly remembered by a figure the official document never used.

What Congress's Own Investigators Could Actually Prove

On April 7, 1992, the House Committee on Government Operations heard testimony on Patriot's real performance from two directions at once. The General Accounting Office's own witness delivered testimony later published as "Operation Desert Storm: Project Manager's Assessment of Patriot Missile's Overall Performance Is Not Supported," concluding that the Army's claims outran its evidence. On the same day, MIT physicist Theodore Postol, who had independently analyzed video of Patriot engagements, told the committee: "The evidence from these preliminary studies indicates that Patriot's intercept rate could be much lower than ten percent, possibly even zero."

GAO's fuller investigation, released that September as "Operation Desert Storm: Data Does Not Exist to Conclusively Say How Well Patriot Performed," found that only 9 percent of Patriot-Scud engagements were supported by the strongest evidence available, physical debris or radar data confirming a Scud warhead had actually been destroyed or disabled. For the rest, GAO concluded the Army could show only that Patriots had detonated near the incoming Scuds, not that they had destroyed them. The Army's own revised methodology, which ranked each engagement's supporting evidence as high, medium, or low confidence before deciding whether it counted as a warhead kill, a lesser mission kill, or a miss, credited Patriot with a warhead kill in 52 percent of engagements at some level of confidence, but the Army itself was highly confident in only about 25 percent of them, a figure still well above the 9 percent GAO would accept as conclusively proven. Raytheon disputed the conclusions publicly, telling PBS's Frontline in a written response that the GAO and Postol analyses relied on incomplete data and understated the system's real performance, a rebuttal that did not change GAO's published findings.

The Scud the Patriot Never Even Saw

The war's deadliest single Scud strike on American forces came on February 25, 1991, three days before the ceasefire, when a missile hit a U.S. Army barracks in Dhahran, Saudi Arabia, killing 28 soldiers. The Patriot battery assigned to defend that site never engaged the incoming missile at all, because it never tracked it as a threat.

GAO's report on the incident, issued nearly a year later on February 11, 1992, traced the failure to a software bug: after running continuously for more than 100 hours, an accumulating error in the system's internal clock threw off its tracking calculations by roughly a third of a second, enough for the battery to lose track of a Scud closing at several times the speed of sound. Army officials already knew about the problem. Israeli data two weeks earlier had shown accuracy degrading after about eight hours of continuous operation, and the Army had a corrected version of the software in hand before the Dhahran attack happened. It arrived at that specific battery on February 26, 1991, the day after the strike it might have prevented. Nearly a century earlier, a Navy court of inquiry took five weeks to blame a mine for a ship explosion in Havana harbor, and two later investigations still couldn't agree with it or with each other.